Sub Pixel Classification Analysis for Hyperspectral Data (Hyperion) for Cairo Region, Egypt

摘要

Traditional hard classifiers in remote sensing applications can label image pixels only with one class, so landcover (e.g. trees) can only be recorded as either present or absent. This approach might lead to inaccurate imageclassification and accordingly inaccurate land cover. The proposed analysis technique provides the relativeabundance of surface materials and the context within a pixel that may be a potential solution to effectivelyidentifying the land-cover distribution. This research is applied on the central region of Cairo using hyperspectralimage data, which provides a large amount of spectral information. A spectral mixture analysis approach is usedon Hyperion data (hyperspectral data) to produce abundance images representing the percentage of the existenceof each material/land cover with a pixel. The uniqueness of this study comes from the fact that it is the first timeHyperion data has been used to extract land cover in Egypt.